The public [scraper catalog](/scrapers) is the authoritative discovery surface for public platforms and capabilities.

## Platform pages

Each platform page is built from its registered module snapshot and can include:

- platform name, category, tagline, maturity, and use cases;
- current health and reliability context;
- capability IDs, names, and descriptions;
- fixed credit cost;
- cURL request examples;
- exact input schema;
- a redacted sample response when publishing one is safe;
- a machine-readable LLM brief.

The web layer does not special-case platform IDs. Fix missing or weak platform presentation in the module manifest and register a new snapshot.

## Machine catalog

Use `/scrapers/llm.md` for a single Markdown reference containing every public platform and capability. Use `/scrapers/:platform/llm.md` when an agent only needs one platform.

## Visibility

Public documentation and machine indexes must never reveal private or unlisted platforms. Authenticated catalog and MCP discovery remain account-scoped and may include private definitions granted to that account.

## Contract ownership

Input schemas, example inputs, timeouts, canaries, and credit cost originate in `module.manifest.json`. Sample responses originate in committed, reviewed fixtures. Public copy comes from `x-catalog` and artwork from `x-ui`.